Head-related transfer functions (HRTFs) characterize sound transmission between specified points in space and the eardrum of a listener. These transfer functions describe how the sound is modified as it propagates along a direction through free space from a given point to the inner ear or eardrum. These modifications are generally caused by the shape of the listener's outer ear as well as the shape and posture of the listener's head and torso. Modelling of HRTFs for use in audio applications allows for more realistic rendering of sounds, particularly the rendering of localized sounds in virtual 3D environments.
HRTFs comprise a collection of transfer functions for each ear, and for each spatial direction (e.g., azimuth and elevation), which encode the frequency response of the listener to an impulsive sound. HRTFs capture the filtering effects of the listener's pinna (outer ear), head and torso. Since each individual's physical properties (anthropometry) differs, their HRTFs will also differ, and thus HRTF measurements need to be performed for each person. Unfortunately, however, acoustic HRTF measurement is a difficult task that is typically performed by trained technicians in a laboratory setting, with the use of specialized and expensive equipment. Additionally, invasive placement of measurement instrumentation in the subject's body is required. For example, traditional acoustic HRTF measurements are performed in an anechoic chamber using a speaker array mounted around individual miniaturized microphones placed inside the individual's ear canals. For these reasons, HRTF-based audio has remained impractical for common use, even in high end audio applications.